Hành vi mỏi do uốn dẻo của dầm bê tông cốt thép được tăng cường bằng vải FRP và hệ thống laminate tiền chế

Flexural Fatigue Behavior of Reinforced Concrete Beams Strengthened with FRP Fabric and Precured Laminate Systems Mahmut Ekenel1; Andrea Rizzo2; John J Myers3; and Antonio Nanni4 Abstract: Rehabilitation of existing structures with carbon fiber reinforced polymers ͑CFRP͒ has been growing in popularity because they offer resistance to corrosion and a high stiffness-to-weight ratio This paper presents the flexural strengthening of seven reinforced concrete ͑RC͒ beams with two FRP systems Two beams were maintained as unstrengthened control samples Three of the RC beams were strengthened with CFRP fabrics, whereas the remaining two were strengthened using FRP precured laminates Glass fiber anchor spikes were applied in one of the CFRP fabric strengthened beams One of the FRP precured laminate strengthened beams was bonded with epoxy adhesive and the other one was attached by using mechanical fasteners Five of the beams were tested under fatigue loading for two million cycles All of the beams survived fatigue testing The results showed that use of anchor spikes in fabric strengthening increase ultimate strength, and mechanical fasteners can be an alternative to epoxy bonded precured laminate systems DOI: 10.1061/͑ASCE͒1090-0268͑2006͒10:5͑433͒ CE Database subject headings: Fatigue; Concrete, reinforced; Concrete beams; Laminates; Rehabilitation Introduction Fiber reinforced polymers ͑FRP͒ have gained importance in bridge rehabilitation in recent years The main reason is their high stiffness-to-weight ratio over steel plates Moreover, these materials are less affected by corrosive environmental conditions, known to provide longer life and require less maintenance A manual lay-up method using two-part epoxies is the conventional way to bond composite fabrics and precured laminates to concrete substrate The main disadvantage of this method is the peeling stresses that may be induced at the location of cracks or ends of the fabric or precured laminates, stresses which tend to pull the strip away from the concrete The peeling of carbon FRP ͑CFRP͒ composite may cause a sudden and catastrophic failure of the structure Another disadvantage is the detachment of CFRP fabric due to the vertical displacement of concrete Graduate Research Assistant, Center for Infrastructure Engineering Studies, Dept of Civil, Architectural, and Environmental Engineering, Univ of Missouri-Rolla, Rolla, MO 65409 E-mail: mex84@umr.edu Graduate Research Assistant, Center for Infrastructure Engineering Studies, Dept of Civil, Architectural, and Environmental Engineering, Univ of Missouri-Rolla, Rolla, MO 65409 E-mail: arx8d@umr.edu Associate Professor, Center for Infrastructure Engineering Studies, Dept of Civil, Architectural, and Environmental Engineering, Univ of Missouri-Rolla, Rolla, MO 65409 E-mail: jmyers@umr.edu Vernon and Marelee Jones Professor, Director-Center for Infrastructure Engineering Studies, Dept of Civil, Architectural, and Environmental Engineering, Univ of Missouri-Rolla, Rolla, MO 65409 E-mail: nanni@umr.edu Note Discussion open until March 1, 2007 Separate discussions must be submitted for individual papers To extend the closing date by one month, a written request must be filed with the ASCE Managing Editor The manuscript for this paper was submitted for review and possible publication on August 31, 2004; approved on September 14, 2005 This paper is part of the Journal of Composites for Construction, Vol 10, No 5, October 1, 2006 ©ASCE, ISSN 1090-0268/2006/5-433–442/$25.00 caused by shear cracking ͑ACI 2002b; Lopez et al 2003͒ Manual lay-up may be labor intensive if it requires significant surface preparation End anchorage might prevent the premature peeling of CFRP fabrics from the concrete substrate In fact, proper anchoring systems may help CFRP precured laminate to develop higher stresses throughout its length ͑Barnes and Mays 1999͒, decreasing stress concentrations and increasing bond strength The use of spikes can increase the flexural capacity of strengthened beams by as much as 35% when compared to strengthened beams without anchor spikes ͑Eshwar et al 2003͒ The same type of glass fiber anchor spikes were also applied on reinforced concrete ͑RC͒ slabs strengthened with a prestressing FRP system, preventing delamination ͑Yu et al 2003͒ A new system that has recently been developed at the Univ of Wisconsin-Madison yielded successful results of FRP precured laminates attachment to concrete ͓mechanically fastened-fiber reinforced polymers ͑MF-FRP͔͒ ͑Lamanna et al 2001a,b,c; Ray et al 2001a,b; Lamanna 2002; Borowicz 2002͒ The installation of the MF-FRP system has proven to be fast and easy, and requires unskilled labor with common hand tools Moreover, the surface preparation can be reduced at the removal of sizeable protrusions such as form lines The efficiency of this rapid-repair strengthening system was demonstrated by rehabilitating an existing bridge and testing two slabs cut from the structure up to failure ͑Borowicz et al 2004͒ The slabs were strengthened with three and five MF-FRP precured laminates, and the ultimate flexural capacity was 25 and 45% higher than the ultimate flexural capacity of the unstrengthened slab, respectively Failure was reached with compression crushing of the concrete for both specimens while the FRP precured laminates were still firmly attached A similar method of strengthening was developed at the Univ of Missouri-Rolla ͑UMR͒ In lieu of pins, concrete wedge bolts and anchors were used The latter was found to be more effective since the presence of hard aggregates in the concrete could damage the thread of the bolts Three off-system bridges in Phelps JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 / 433 Table Material Properties and Reinforcement Geometry Material Modulus of elasticity ͑GPa͒ Strength ͑MPa͒ Ultimate strain ͑␮␧͒ Thickness ͑mm͒ Width ͑mm͒ Concretea 27.6b 25.1c 3,000 NA NA d e 413.7 200e 2,070 NA NA Reinforcing steel 3792 227.5 16,670 0.165 203.2 CFRP fabricd,f 835g 62g 14,000 3.175 101.6 CFRP precured laminated a Compressive properties b The compressive strength was determined at 28 days according to ASTM C 39-01 ͑2001͒ c The modulus of elasticity in compression was determined at 28 days according to ASTM C 469-94 ͑1994͒ d Tensile properties e The yield strength and the modulus of elasticity were determined according to ASTM A 370-02 ͑2002͒ f Data obtained by the manufacturer g The stress at failure and the modulus of elasticity of the laminate were determined according to ASTM D 3039 ͑2006͒ County, Mo., were strengthened using the MF-FRP system in 2004 ͑Rizzo 2005͒ Thus it was possible to compensate for an insufficient amount of longitudinal reinforcement, in this case the reason for which the entire superstructure elements were visibly cracked Another issue that should be addressed is the performance of these systems under fatigue loading Even though the static behavior of various FRP application systems has been widely studied, fatigue resistance still requires further investigation Composites are believed to have a higher resistance to fatigue compared to other engineering materials ͑Ekenel 2004͒ It has been reported that the fatigue durability of RC beams significantly improved after strengthening with externally bonded FRP laminates ͑Barnes and Mays 1999; Shahawy and Beitelman 1999; El-Tawil et al 2001; Senthilnath et al 2001; Papakonstantinou 2001; Brena et al 2002͒ Previous researchers stated that the fatigue failure of RC beams is not always by the same mechanism as static failure ͑Barnes and Mays 1999͒ An improvement in fatigue life after strengthening with FRP laminates is expected as the increase in stiffness and strength reduces the crack propagation, causing a reduction of stress in the reinforcing steel Grace ͑2004͒ applied a fatigue loading at service load levels for two million cycles on RC beams strengthened with CFRP precured laminate and fabric; no significant effect on the ultimate load-carrying capacity of such beams was observed This study investigates the affect of fatigue loading on flexural residual capacity of two different FRP strengthening techniques: CFRP fabric and precured laminates bonded with epoxy, and FRP precured laminates mounted with mechanical fasteners ͑MF͒ Glass fiber anchor spikes were used as a supplemental end Cross-section area ͑mm2͒ NA NA 33.5 322.6 anchorage for one of the CFRP fabric strengthened beams in order to investigate their efficiency under cyclic and static flexural loading Experimental Investigation Material Properties Table summarizes the material properties for concrete, steel reinforcement, CFRP sheets, and precured laminates It is noteworthy to mention that the CFRP sheet properties are fiber related, whereas the CFRP precured laminates properties are gross cross section related Analytical Design The strength models used to predict the ultimate capacity of the critical section utilize strain compatibility in the section, equilibrium, and constitutive relations of the materials Bonded CFRP fabrics and precured laminate were idealized as linear up until failure In order to avoid cover delamination or FRP debonding, a limitation was placed on the strain level developed in the laminate using the bond-dependent coefficient for flexure ͑␬m͒ according to ACI Committee 440.2R-02 ͑ACI 2002b͒ Table summarizes the design stresses f fu and strains ␧ fu, the coefficient ␬m, the axial stiffness AE and the maximum force A␬m f fu that can be developed by the strengthening of each beam according to ACI Committee 440.2R-02 ͑ACI 2002b͒ Coefficient ␬m was set equal to one for the beam strengthened with FRP fabrics using anchor spikes In reality, the presence of the anchor spikes reduces the probability of having cover delamination or FRP debonding at the Table Parameters Used for Design ␧ fu ͑␮␧͒ AE ͑kN͒ A␬m f fu ͑kN͒ 7,633 7,633 114 127 460 7,345 0.551 20,190 148 531 8,483 0.636 18,090 171 Strengthening description f fu ͑MPa͒ CFRP fabric CFRP fabric with anchor spikes CFRP precured laminate with epoxy ͑bonded SafStrip͒ CFRP precured laminate with wedge anchors ͑bolted SafStrip͒ 3,385 3,792 ␬m 14,880 0.893 16,670 1.000 Fig Details of the connection between concrete and CFRP precured laminate with wedge anchors ͑mm͒ 434 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 Table Beams Cross Section Properties Strengthening description Fig Shear-bond test setup ends of the fabric by absorbing part of the significant stress concentration ͑Eshwar et al 2003; Smith and Teng 2001͒ In order to simplify the calculation of the section properties by using the same model valid for bonded FRP systems, a different definition for the coefficient ␬m was introduced for FRP pre-cured connections to take into account the different mechanisms involved in the stress transferring process of this type of strengthening ͑such as net tension failure at open hole, concrete substrate failure, number and pattern of the fasteners, clamping pressure, concrete substrate failure, etc.͒ In addition, due to the particular behavior of the connection as described in the next paragraph, it was possible to estimate the coefficient ␬m,bolted as ␬m,bolted = min͑FFRP,nfasteners · Pbearing͒ AFRPEFRP ␧ fu where ␧ fu, AFRP, and EFRPϭstrain at failure of the laminate, the cross-section area, and modulus of elasticity, respectively; FFRPϭmaximum allowable load in the laminate corresponding to the net section and nfasteners is the number of fasteners on one side of the beam; Pbearingϭbearing capacity of the connection which takes the mode of failure of the connection into account depending on its geometrical details and clamping pressure The previous expression of ␬m,bolted, valid for this particular MF-FRP system, needs to be reformulated and validated in further research works to obtain a more general formula Load at cracking ͑kN͒ Load at yielding ͑kN͒ Unstrengthened 9.8 25.5 CFRP fabric * 35.5 CFRP fabric with * 35.5 anchor spikes CFRP precured laminate * 51.5 with epoxy ͑bonded SafStrip͒ CFRP precured laminate * 51.5 with wedge anchors ͑bolted SafStrip͒ Note: *ϭthese beams were already cracked Load at failure ͑kN͒ Expected mode of failure 28.9 57.5 57.5 Compression Compression Compression 81.1 Tension ͑Peelingdelamination͒ Compression 82.3 In order to mechanically fasten the FRP laminate to the concrete, the optimal solution in terms of mechanical behavior of the connection was determined as a result of an experimental program conducted at UMR ͑Rizzo 2005͒ The results of the bearing tests showed that the bearing capacity of the laminates is proportional to the diameter of the pin and the area restrained by the washer, which avoids the buckling of the fibers that are directly in contact with the pin On the other hand, the bearing capacity can be reached only if the distance of the hole from the free edge is higher than 50.8 mm, thus preventing other failure modes such as shear-out or cleavage The fastening system used in this experimental campaign can be seen in Fig The diameter of the pin was chosen to avoid the failure of the connection in the concrete substrate The load that causes spalling of the concrete was calculated according to the ACI 355.1R-91 document ͑ACI 1991͒ The minimum embedment depth of the bolts was designed according to the recommendations provided by the manufacturer Single-bolted shear-bond tests ͑Fig 2͒ were performed in order to obtain the ultimate shear capacity of the connection It was observed that, for concrete having an f Јc = 27.5 MPa, the failure mode was due to bearing of the FRP at 14.0 kN Multibolted specimens were also tested to understand the stress distribution between the fasteners, and it was found that for the present CFRP Fig Geometry of the specimens: ͑a͒ Longitudinal section ͑mm͒; ͑b͒ cross section ͑mm͒ JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 / 435 Fig Test plan for anchor spikes placement on beam ͑mm͒ precured laminates, the applied load can be assumed as uniformly distributed between all the fasteners at the ultimate conditions ͑Rizzo 2005͒ Therefore, for the specimen strengthened with MF-FRP system, 22 evenly spaced fasteners were used to attach the precured laminate in the shear spans in order to induce first concrete failure followed by the bearing failure at the connections The spacing between the fasteners was chosen to satisfy the recommendations provided by ACI 355.1R-91 ͑ACI 1991͒ document and the manufacturer of the anchors Although the FRP precured laminate by itself is a linear elastic material, the MF-FRP precured laminate shows plastic behavior beyond the bearing strength Pb In reality, it is possible to obtain a pseudoplastic behavior of the connection by using the proper geometrical details ͑edge distance, washer, gap filler, etc.͒ Herein, the term “plastic” is used in the sense that the maximum load reached in the connection is not abruptly followed by a drop but it can be maintained until large elongation of the hole Therefore, the behavior of the MF-FRP connection was modeled as elastic-perfectly plastic The strain distribution across the section was calculated assuming that the strains in the reinforcement and concrete are directly proportional to the distance from the neutral axis, that is, a plane section before loading remains plane after loading In the case of the MF-FRP precured laminate, this assumption is not completely accurate because there is not intimate contact between the concrete and external FRP reinforcement This approach was used for convenience of calculation as an approximation to avoid the relative slip between the MF-FRP laminate and the concrete substrate due to the presence of gaps in the connection components, the elastic deformation of the bolts, and the bearing mechanisms corresponding to the holes’ locations It is important to underline that the load distribution for the MF-FRP precured laminate under bending is complex It is understood that bearing of the FRP precured laminate at the location of the fasteners introduces nonuniformity to the strain across the section In the context of this work, a uniform distribution of stress across the section of FRP precured laminate was assumed as a simplification of the problem This approach allows us to estimate the resultant tension force in the FRP strengthening as AFRPEFRP␧FRP, where AFRP, EFRP, and ␧FRP are the cross section, the modulus of elasticity, and the strain of the FRP laminate, respectively ͑␧FRP is calculated in the previous assumption that plane cross sections before loading remain plane after loading͒ ͑Rizzo 2005͒ Table summarizes the design properties of the cross section for all the beams The expected failure loads for the beams strengthened with bonded fabric sheets are the same because the expected modes of failure are crushing of the concrete for both of them even though the ␬m factors are different It is also noteworthy to mention that the expected mode of failure for the beam with the bonded CFRP precured laminate is peelingdelamination; whereas the one strengthened with the MF-FRP system is compression Sample Preparation Seven RC beams were fabricated in the laboratory for this investigation The width and the height of the cross section are 254 and 165 mm, respectively The total area and the effective depth of the tension side steel reinforcement are 214 mm2 and 122 mm, respectively The total area and the clear concrete cover of the compression side steel reinforcement are 142 mm2 and 45 mm, respectively The dimensions and cross-sectional details of the Fig Beam strengthened with bonded CFRP fabrics and using anchor spikes: ͑a͒ Half saturated dry glass fibers; ͑b͒ final shape of anchor spikes 436 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 Fig Beam strengthened with bonded CFRP precured laminate ͑epoxy application͒ test beams are shown in Fig Two beams served as unstrengthened control specimens ͑S-0 and S-0F͒ Beam S-0 was tested under static loading while beam S-0F cycled under million fatigue cycles prior to static testing ͑F represents fatigue cycling͒ All of the beams were precracked before strengthening by loading the beams beyond the cracking load to simulate the condition of a typical RC beam prior to repair/strengthening Three beams were strengthened with a single CFRP ply throughout the length of the tension face of test specimens ͑S-1, S-1F, and S-2F͒ Except cleaning by wire brushing and pressurized air, no other surface preparation method was applied in order to simulate a worse case situation for bond prior to cycling The CFRP fabric was applied in approximately 45 per beam The curing time of epoxy was 48 h ͑under laboratory conditions͒ as recommended by manufacturer Four anchor spikes were applied on one of the CFRP fabric strengthened beams ͑Beam S-2F͒ The anchors were located at the ends of the sheet, where high peeling and shear stresses may develop The purpose was to prevent the premature peeling of CFRP laminates by anchoring CFRP fabrics to the concrete The locations were determined based on the information obtained from the previous studies done by Sagawa et al ͑2000͒, Eshwar et al ͑2003͒, and Yu et al ͑2003͒ The strengthening plan is illustrated in Fig Spikes were formed from dry glass fibers, and they were half dry and half coated with epoxy Glass fibers were preferred because of their economical advantages The epoxycoated part had a diameter of approx 9.5± 1.5 mm ͓see Fig 5͑a͔͒ First, four holes were drilled into the concrete with the dimensions of 25.4 mm in depth and 12.7 mm in diameter Then, primer was applied on the tension side of the beam, which was followed by the application of saturant The holes were also filled with saturant to the half of their depths After applying the CFRP fabric, the precured part of the anchor spikes was inserted into the holes throughout the fabric The dry fibers were fanned out over the CFRP fabric Finally, a second layer of saturant was applied and roller spikes used to avoid any air bubbles at the interface ͓Fig 5͑b͔͒ The system was cured for 48 h in laboratory environment according to manufacturer’s recommendation Two of the beams were strengthened using CFRP precured laminates ͑S-3F and S-4F͒ For beam S-3F, the precured laminate was bonded over the tension zone of the concrete using epoxy The epoxy was laid on the concrete with a thickness of 3.2 mm ͑see Fig 6͒ Finally the CFRP precured laminate was placed over epoxy and pressure was applied by a spike roller to force air bubbles out The application was done in approximately 15 min/ beam For Beam S-4F, the precured laminate was me- Fig Beam strengthened with mechanically fastened CFRP precured laminate chanically fastened to the concrete surface by means of concrete wedge anchors with 9.5 mm diameter and 57.1 mm total length First, holes were drilled according to the design pattern into the concrete to a depth of 50.8 mm using a 9.525 mm diameter solid carbide-tipped bit The precured laminate was drilled using the same bit and pattern of holes, cleaned of dust, and laid on the concrete surface Finally, the fasteners were hammered into the holes over the CFRP precured laminate making sure that the nut and the washer were resting solidly against the fixture The steel washer had an 11.1 mm inner diameter, a 25.4 mm outer diameter, and a 20.6 mm thickness At this point, the gaps between anchors and FRP were filled with epoxy in a way that the amount of resin was enough to wet the washer-FRP interface In this fashion, the rigid movement of the anchors was not prevented as the gaps between the fasteners and the concrete sleeve were not filled Finally, the nuts were tightened with a wrench to a torque in the range of 34.0– 41.0 N m ͑Fig 7͒ according to the manufacturer’s specifications in order to limit the shear stress due to the torsion on the anchor The application took about one hour The test plan is presented in Fig Test Setup and Procedure Beams S-0F, S-1F, S-2F, and S-4F were tested for fatigue over a simply supported span of 1,829 mm The beams were loaded with two concentrated loads placed at equal distance ͑152 mm͒ from the beam centerline The supports were placed 63 mm away from the end points A sketch of the test setup is shown in Fig 3͑a͒ All five beams were cycled under fatigue loading between a minimum of 33% and a maximum of 63% of the theoretical ultimate flexural capacity of the section These load percentages were extrapolated based on the safety factors proposed by AASHTO ͑2000͒ which were simulated common service conditions that a structure, like a bridge, might experience during its lifetime Table shows the applied minimum and maximum loads for each beam Subjecting the beams to the same percentage of the ultimate moment capacity seemed an appropriate method to compare fatigue performance of fabrics strengthened and precured laminate strengthened beams with different capacity Because the maximum load of Beams S-3F and S-4F exceed the yielding loads presented at Table 3, it is noteworthy to mention that the values in Table are theoretical calculations The tests were terminated either when the beam failed or reached million cycles, whichever occurred first The frequency applied during the fatigue testing was cycles/ s ͑2 Hz͒ Prior to actual fatigue testing, a ten-cycle test was performed JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 / 437 Fig Concrete wedge anchors pattern ͑mm͒ to measure the mid-span displacement of the beams by using linear variable differential transducer ͑LVDT͒ The frequency was adjusted to 0.1 Hz during this ten-cycle test in order to capture adequate number of data points The same tests were performed during the fatigue testing after every 0.5, 1, 1.5, and million cycles and the corresponding stiffness was determined Experimental Results Fatigue Test Results All beams survived million cycles The initial and million cycle stiffness measurements for the unstrengthened control sample ͑S-0F͒ were 7.46 and 6.30 kN/ mm, respectively ͑see Fig 9͒ This corresponds to a decrease of 16% at million cycles as compared to initial cycle The stiffness mentioned herein is defined as the slope of the load versus midspan displacement relation between the maximum and minimum specified loads Fig shows the measured stiffness versus number of cycles Fig 10 shows the midspan deflections versus number of cycles As shown in Fig 9, it was observed that most of the stiffness loss occurred between first and 0.5 million cycles The phenomenon is related to the opening and propagation of the cracks, which implies a relative slip between concrete, steel reinforcement, and the FRP strengthening ͑local microdebonding at the tips of the cracks in the case of bonded systems, partial rotation of the fasteners, relative slip between the components of the connections due to the presence of gaps and local microcrushing of the concrete in the contact area in the case of bolted system͒ as well as failure of Fig Stiffness versus number of cycles Table Fatigue Test Program Beam number S-0F S-1F S-2F S-3F S-4F Strengthening description Minimum and maximum load Applied ͑kN͒ Unstrengthened CFRP fabric CFRP fabric with anchor spikes CFRP precured laminate with epoxy CFRP precured laminate with wedge anchors 9.8–18.7 17.3–33.4 17.3–33.4 28.9–55.2 28.9–55.2 Fig 10 Maximum deflections versus number of cycles 438 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 Fig 12 Load versus midspan curves up to proportional limit for all fatigued beams the concrete in the tensile zone ͑not accounted in conventional calculations͒ After million cycles, all the previous microfailures reduced The interlocking of the steel bars ͑due to the ribs on the surface͒ avoided further relative slip at the interface steel concrete; the debonding zones and cracks in the concrete stabilized and the fasteners were fixed in a stable position The highest change in stiffness was observed in CFRP pre-cured laminate strengthening with mechanical fasteners ͑S-4F͒ by 22% at million cycles as compared to initial cycle All others ͑S-1F, S-2F, and S-3F͒ showed a decrease of 15% on average at million cycles Flexural Test Results Fig 11 Load versus midspan displacement curves until test beams failure ͑Fϭfatigued prior to static test͒: ͑a͒ Raw data; ͑b͒ normalized data All seven beams were tested under flexural loading The flexural tests were performed under four point bending and mid-span displacements were recorded using LVDT Fig 11͑a͒ exhibits the load versus midspan displacement diagrams of test beams loaded until failure As shown in Fig 11͑a͒, Beam S-2F exhibited a 39% higher capacity as compared to Beam S-1 The increase in capacity can be attributed to the presence of the anchor spikes that improved the bond properties at the concrete-CFRP interface Table Expected and Measured Failure Loads Failure load Beam number Strengthening description S-0 S-0F S-1 S-1F S-2F S-3F S-4F a Measured failure Unstrengthened Unstrengthened CFRP fabric CFRP fabric CFRP fabric with anchor spikes CFRP precured laminate with epoxy CFRP precured laminate with wedge anchors load/expected failure load E sA s + E f A f E sA s Expected ͑kN͒ Measured ͑kN͒ Normalizeda ͑kN͒ 1.000 1.000 1.179 1.179 1.179 1.449 1.449 28.9 28.9 57.5 57.5 57.5 81.1 82.3 33.0 33.4 64.1 63.2 89.5 80.4 71.1 33.0 33.4 54.4 53.6 75.9 55.5 49.1 JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 / 439 Fig 13 Modes of failure of the specimens: ͑a͒ Unstrengthened control sample; ͑b͒ CFRP fabric; ͑c͒ CFRP fabric with anchor spikes; ͑d͒ bonded FRP precured laminate ͑bonded SafStrip͒; ͑e͒ MF-FRP system ͑bolted SafStrip͒ This can be seen in the failure modes between S-1 and S-2F S-1 failed by complete CFRP debonding; whereas, anchor spikes were still holding until S-2F failed by delamination/fracture of fabric Table compares the expected and the measured failure loads The measured loads were also normalized in order to provide a fair analysis between the different strengthening techniques ͓see Fig 11͑b͔͒ The normalization was performed by dividing the measured loads by a stiffness coefficient, defined as the ratio between the total tension strengthening ͑CFRP+ steel͒ stiffness and the steel reinforcement one As shown in Table 5, the average increase in ultimate failure load by single-ply CFRP strengthening as compared to unstrengthened beam was 94% Beams S-2F and S-3F exhibited failure loads which were 39 and 2% higher than S-1, respectively; however, S-4F showed a failure load which was 10% lower than S-1 The relative slip between concrete and precured laminate that might occur in the case of the beam strengthened using the MF-FRP system ͑S-4F͒ explains the lower value of strength found as compared with the beams strengthened with the bonded system In reality, the engagement between fasteners and FRP precured laminate was complete after steel yielded This phenomenon allowed for a larger magnitude of crack propagation The effect can be attributed to the accumulation of the damages at the location of the fasteners In fact, due to the particular pattern of the fasteners and loading configuration, six anchors in the midspan of the specimen ͑three for each side͒ experienced a load higher than the average load at bearing of the single connection during the fatigue cycles Consequently, the effective fasteners that anchored the laminate were only 16, not 22 Recalculating with only 16 fasteners, the ultimate load at failure results 74.8 kN, value very close to the experimental one ͑71.1 kN͒ This conclusion underlines the importance to check the connections of a MF-FRP fastening system for both ultimate and service load conditions ͑Rizzo 2005͒ Occasional overloads can highly affect the performance of the strengthening at ultimate condition A decrease in ductility due to fatigue cycling can also be seen in Fig 11 Beams S-0F and S-1F displayed midspan deflections which were 13 and 18% lower than S-0 and S-1, respectively However, there was no decrease in ultimate strength The decrease in ductility can be explained by the higher crack formation due to the fatigue cycles 440 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 Fig 12 shows the load versus midspan displacement curves of fatigued beams up to the proportional limit As illustrated, the slopes of the load-deflection curves measured in static loading ͑see Fig 12͒ are lower than the ones measured in fatigue loading ͑see Fig 9͒ This can be speculated as the hysteresis effect which caused the RC beam to start a successive cycle before being fully recovered from the previous one; however, this parameter can be used for comparison purposes among each representative specimen As shown in Fig 12, Beams S-2F, S-3F, and S-4F exhibited slopes which were 48, 22, and 3% higher than S-1, respectively Fig 13 shows the failure modes of the test samples Both unstrengthened control samples ͑S-0 and S-0F͒ failed by concrete crushing The CFRP fabric applied samples ͑S-1 and S-1F͒ failed by concrete crushing followed by a complete CFRP delamination As a consequence of the extended cracking phenomena, CFRP fabric was detached and torn off ͓see Fig 13͑b͔͒ The failure mode of Beam S-2F was crushing of the concrete Analyzing the specimen after failure, it was possible to detect that the failure mode of Beam S-2F was crushing of the concrete, followed by delamination of the CFRP fabric at the midsection; consequently, breaking on one side close to the anchor spikes Both beams strengthened with CFRP fabrics exhibited very similar midspan displacement readings at failure load and both failures were instant and catastrophic The beam strengthened with FRP precured laminate bonded with epoxy ͑S-3F͒ failed at a load of 13% higher than the beam strengthened with MF-FRP system ͑S-4F͒ The failure of the Beam S-3F was catastrophic After the compression crushing of the concrete, the precured laminate peeled off at one side of the beam ͓see Fig 13͑d͔͒ On the other hand, the failure of the Beam S-4F was very ductile; the pre-cured laminate was firmly attached at the surface of the concrete until very large deflections occurred with rotation of the majority of fasteners ͓see Fig 13͑e͔͒ After the test, the beam was rolled upside down allowing the detection of extensive bearing failure at the locations of the fasteners Conclusions The following conclusions can be drawn based on the fatigue and flexural tests results presented in this paper: • The FRP strengthening increased the fatigue life of RC beams by increasing stiffness and reducing crack propagation; • The change in stiffness at million cycles as compared to initial cycle was approximately the same for all the beams ͑15%͒ except Beam S-4F ͑22%͒ This is due to the fasteners which allowed greater crack formation and propagation until the complete engagement of the strengthening; • The fatigue loading slightly reduced the members’ ductility but did not significantly affect the failure load for the CFRP fabric strengthened and unstrengthened specimens tested; • Based on the flexural test results, it can be concluded that the analytical design using ACI Committee 440.2R-02 ͑ACI 2002b͒ was conservative in calculating the ultimate capacity of the beam even after million fatigue-cycling at service load, except for the beam strengthened with mechanically fastened FRP precured laminate which showed a load at failure 14% lower than the expected value This can be partially attributed to the higher damage accumulation in the FRP precured laminate around the anchorage holes Monitoring the damage accumulation in cyclic loading would be interesting for future investigation; • The fatigue and static loading exhibited that the use of me- chanical fasteners can be an alternative to the epoxy bonded systems Moreover, the beam strengthened with MF-FRP showed a more desirable apparent ductile behavior as compared to the beam strengthened with epoxy bonded FRP system The increase in ductility exhibited by Beam S-4F was 3.5 times that of Beam S-3F; • The use of anchor spikes ͑S-2F͒ resulted in a significant increase ͑39%͒ in the ultimate capacity of the beam as compared to CFRP strengthened beam without anchor spikes ͑S-1͒ The increase in capacity can be attributed to the presence of the anchor spikes that improved the bond properties at the concrete-CFRP interface, which can be seen in the failure modes of Beams S-1 and S-2F The increase in labor costs using this anchorage technique could be offset by a reduction in the flexural reinforcement used Acknowledgments The writers wish to express their gratitude and sincere appreciation to the authority of Federal Highway Administration ͑FHwA͒ and the Center for Infrastructure Engineering Studies ͑CIES͒ at the University of Missouri-Rolla ͑UMR͒ for supporting this research study They would like to thank Larry Bank from the University of Wisconsin and Nestore Galati and Jason Cox for their contribution to this research The authors would also like to acknowledge Nathan Marshall and Jared Brewe for their effort as undergraduate research assistants References American Association of State Highway and Transportation Officials ͑AASHTO͒ ͑2002͒ Standard specifications for highway bridges, 17th Ed., Washington, D.C American Concrete Institute ͑ACI͒ ͑1991͒ “State-of-the-art report on anchorage to concrete.” ACI Committee 355.IR-91, Detroit American Concrete Institute ͑ACI͒ ͑2002b͒ “Guide for the design and construction of externally bonded FRP systems for strengthening concrete structures.” ACI Committee 440.2R-02, Farmington Hills, Mich ASTM ͑1994͒ “Standard test method for static modulus of elasticity and Poisson’s ratio of concrete in compression.” ASTM C 469, Vol 04-02, Philadelphia ASTM ͑2001͒ “Standard test method for compressive strength of cylindrical concrete specimens.” ASTM C 39, Vol 04-02, Philadelphia ASTM ͑2002͒ “Standard test methods and definitions for mechanical testing of steel products.” ASTM A 370, Vol 04-02, Philadelphia ASTM ͑2006͒ “Standard test method for tensile properties of polymer matrix composite materials.” ASTM D 3039, Vol 15-02, Philadelphia Barnes, R A., and Mays, G C ͑1999͒ “Fatigue performance of concrete beams strengthened with CFRP plates.” J Compos Constr., 3͑2͒, 63–72 Borowicz, D T ͑2002͒ “Rapid strengthening of concrete beams with powder actuated fasteners and fiber reinforced polymer ͑FRP͒ composite materials,” M.S thesis, Univ of Wisconsin–Madison, Madison, Wis Borowicz, D T., Bank, L C., Nanni, A., Arora, D., Deza, U., and Rizzo, A ͑2004͒ “Ultimate load testing and performance of bridge strengthened with fiber reinforced composite materials and powder-actuated fasteners.” Proc., 83rd Annual Transportation Research Board Meeting ͑CD-ROM͒, Washington, D.C Brena, S F., Wood, S L., and Kreger, M E ͑2002͒ “Fatigue tests of reinforced concrete beams strengthened using carbon fiber reinforced polymer composites.” Proc., 2nd Int Conf on Durability of FRP Composites for Construction, Montreal, Quebec, 575–586 JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 / 441 Ekenel, M ͑2004͒ “Durability performance of advanced construction materials used in infrastructure repair and rehabilitation systems.” Ph.D dissertation, Univ of Missouri-Rolla, Rolla, Mo El-Tawil, S., Ogunc, C., Okeil, A., and Shahawy, M ͑2001͒ “Static and fatigue analyses of RC beams strengthened with CFRP laminates.” J Compos Constr., 5͑4͒, 258–267 Eshwar, N., Ibell, T., and Nanni, A ͑2003͒ “CFRP strengthening of concrete bridges with curved soffits.” Proc., Int Conf on Structural Faults and Repairs, Montreal, Quebec Grace, N F ͑2004͒ “Concrete repair with CFRP.” Concr Int., 26͑5͒, 45–52 Lamanna, A J ͑2002͒ “Flexural strengthening of reinforced concrete beams with mechanically fastened fiber—reinforced polymer strips.” Ph.D thesis, Univ of Wisconsin-Madison, Madison, Wis Lamanna, A J., Bank, L C., and Scott, D W ͑2001a͒ “Flexural strengthening of RC beams using fasteners and FRP strips.” ACI Struct J., 98͑3͒, 368–376 Lamanna, A J., Bank, L C., and Scott, D W ͑2001b͒ “Rapid flexural strengthening of RC beams using powder actuated fasteners and FRP strips.” Proc., FRPRCS-5 Fiber Reinforced Plastics for Reinforced Concrete Structures, Vol 1, Univ of Cambridge, Cambridge, U.K., 389–397 Lamanna, A J., Bank, L C., and Scott, D W ͑2001c͒ “Rapid strengthening of RC beams using powder actuated fasteners and FRP strips.” Proc., 5th Int Symp of FRP in Reinforced Concrete Structures, Cambridge, U.K., 389–397 Lopez, M M., Naaman, A E., Pinkerton, L., and Till, R D ͑2003͒ “Behavior of RC beams strengthened with FRP laminates and tested under cyclic loading at low temperature.” Int J Mater Prod Technol., 19͑1–2͒, 108–117 Papakonstantinou, C G., Petrou, M F., and Harries, K A ͑2001͒ “Fatigue behavior of RC beams strengthened with GFRP sheets.” J Compos Constr., 5͑4͒, 246–253 Ray, J C., Scott, D W., Lamanna, A J., and Bank, L C ͑2001a͒ “Flexural behavior of reinforced concrete members strengthened using mechanically fastened fiber reinforced polymer plates.” Proc., U.S Army Science Conf., Baltimore Ray, J C., Velazquez, G I., Lamanna, A J., and Bank, L C ͑2001b͒ “Rapidly installed fiber-reinforced polymer ͑FRP͒ plates for upgrade of reinforced concrete bridges.” High Performance Materials in Bridges and Buildings, Koha, Hawaii Rizzo, A ͑2005͒ “Application in off-system bridges of mechanically fastened FRP ͑MF-FRP͒ procured laminates.” M.Sc thesis, Univ of Missouri-Rolla, Rolla, Mo Sagawa, Y., Matsushita, H., Takeo, K., and Saito, M ͑2000͒ “A study on a method of anchorage of carbon sheet for flexural strengthening.” Trans Jpn Concr Inst., 22, 237–242 Senthilnath, P S., Belarbi, A., and Myers, J J ͑2001͒ “Performance of CFRP strengthened RC beams in the presence of delaminations and lap splices under fatigue loading.” Proc., Composites in Construction: 2001 Int Conf., Porto, Portugal Shahawy, M., and Beitelman, T E ͑1999͒ “Static and fatigue performance of RC beams strengthened with CFRP laminates.” J Struct Eng., 125͑6͒, 613–621 Smith, S T., and Teng, J G ͑2001͒ “Interfacial stresses in plated beams.” Eng Struct., 23, 857–871 Yu, P., Silva, P F., and Nanni, A ͑2003͒ “Flexural strengthening of concrete slabs by a three-stage prestressing FRP system enhanced with presence of GFRP anchor spikes.” Proc., Composites in Construction Int Conf., Rende ͑CS͒, Italy 442 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / SEPTEMBER/OCTOBER 2006 View publication stats ... of FRP precured laminate was assumed as a simplification of the problem This approach allows us to estimate the resultant tension force in the FRP strengthening as AFRPEFRP? ?FRP, where AFRP, EFRP,... yielding ͑kN͒ Unstrengthened 9.8 25.5 CFRP fabric * 35.5 CFRP fabric with * 35.5 anchor spikes CFRP precured laminate * 51.5 with epoxy ͑bonded SafStrip͒ CFRP precured laminate * 51.5 with wedge anchors... recommendations provided by ACI 355.1R-91 ͑ACI 1991͒ document and the manufacturer of the anchors Although the FRP precured laminate by itself is a linear elastic material, the MF -FRP precured laminate